Numerical modeling is the process of obtaining an objective forecast of the future state of the atmosphere by solving a set of equations that describe the evolution of variables (temperature, wind speed, humidity, pressure) that define the state of the atmosphere.

The process begins with analysing the current state of the atmosphere by taking a previous short-range forecast and using observations to amend this forecast so that the best of the current true state of the atmosphere is obtained. All numerical models of the atmosphere are based upon a set of governing equations. Numerical models differ in the approximations and assumptions made in the application of these equations, how they are solved and also in the representation of physical processes.

The simulation of atmospheric processes on a computer with the aim of taking their latest state to derive a prognosis of the future development is called numerical weather forecasting. Nowadays most of the weather forecasts are compiled on this basis.

Global General Circulation Models

Global General Circulation Models (GCMs) cater the needs to understand dynamics of the atmospheric processes which are the long term indicators of the weather with sufficient lead time. As GCMs cover all the globe and high speed Super Computing Systems are involved to process data for predictions at the coarse resolution. GCMs generally run at 100 – 300 km horizontal grid resolution. Therefore, their output can not be used as forecast. However, it can provide sufficient outlook of weather pattern on a broader scale in a particular region. In this hi-tech world, GCMs have also been developing to produce output as finest as the statistical models can do due to the simultaneous progress in processing speed of supercomputers. Since 2010, GCM20 has been leading the world for its finest resolution on horizontal and vertical scales as well as the processing time steps. It was developed by Japanese Scientists at Yokohama Earth Simulator. This model runs at 20km×20km horizontal grid net incorporating all the advanced characteristics of numerical weather prediction.

To produce city scale prediction on extended time scale, however, Regional climate Models and statistical Downscaling Models are commonly used. Pakistan Meteorological Department has been using Regional Climate Model (RegCM4) fourth version developed by ICTP, Trieste, Italy and Précis introduced b Hadley Center, UK for climate outlooks. Regional predictions of monthly and seasonal time scales are prepared by downscaling GCM outputs statistically using SDSM. These techniques are being tested against accuracy of models and simultaneously improvements and alterations are adopted to minimize the errors. Hence Seasonal and Monthly predictions are in experimental phase but regularly uploaded on PMD website for comments and suggestions of users for their improvement.

High Resolution Regional Model (HRM)

Pakistan Meteorological Department (PMD) is using High Resolution Model (HRM) with horizontal resolution of 11km. The HRM is developed by Deutscher Wetterdienst (DWD) National Meteorological Service of Germany. The data of Global Model (GME) also of DWD is used as initial conditions in HRM for the numerical weather forecasts. The use of this high resolution model is helping to improve the weather forecasts issued by PMD and also to enhance the accuracy of weather predictions.

The Global Forecast System (GFS) is a global numerical weather prediction computer model run by NOAA. It is a common program for developing the extended forecast (3 to 10 days out).The trough / ridge pattern can be used as a medium range-forecasting tool. If the GFS model has a good handle on the evolution of the trough / ridge pattern, the forecast will do reasonably well. Keep in mind that beyond about 5 days, the GFS data output can change quite dramatically from one model run to the other. Small changes in the initial input into the model result in dramatic difference beyond the 5 day forecast period (i.e. butterfly effect). The trough / ridge pattern determines which regions will be cooler or warmer than normal and which regions have a higher likelihood of precipitation. Temperatures tend to be cooler than normal under amplified troughs and warmer than normal under amplified ridges. Precipitation tends to occur to the right of a highly amplified trough axis, along fronts and along shortwaves.

GSM is a global model of China meteorological Agency. GSM provide short-and medium range forecast,using global forecast domain contain horizontal resolution of almost 30km. It provides up to one week forecast on 12-hourly basis. The model output is updated daily at 00 UTC.

Climate models are increasingly used for downscaling climate scenarios, seasonal climate predictions, land use climate impact and ocean-atmospheric interaction studies all over the world. A climate model consists of complex computer program / programs. These programs look at several mathematical equations that govern the atmospheric processes at once. They take into account conservation of mass, energy, momentum in a grid box system (the world is divided in to several grid boxes). The model focuses on each grid box and the transfer of energy between grid boxes. The aim is to reproduce as faithfully as possible the real climate system.

Pakistan Meteorological Department (PMD) is using the PRECIS regional climate modeling system of developed by Hadley Centre of Meteorological Service, United Kingdom and RegCM4 (Regional Climate Model version 4)  of ICTP, Trieste, Italy, for developing climate scenarios, climate predictions, land use / land cover climate impact and for climate and climate change related research.  The models that are functional and installed at Pentium4 Servers at Research & Development (R & D) Division of PMD, Islamabad.

Providing REgional Climates for Impacts Studies (PRECIS)

PRECIS is a regional climate model (RCM) ported to run on a Linux PC with a simple user interface, so that experiments can easily be set up over any region of the globe. PRECIS is designed for researchers (with a focus on developing countries) to construct high-resolution climate change scenarios for their region of interest.

Distribution of Dust Storms frequency over Pakistan Annual Normal (1961-1990) 

Distribution of Dust Storms frequency over Pakistan Annual Mean (1991-2000)

Mean Annual Thunderstorms (1991-2000) 

Normal Annual Thunderstorms (1961-1990)